This section describes the so called {\it dynamic arrays}. This is a C
array-like data structure i.e. the member access time is constant (and not
-linear in number of container elements as for linked lists). However, these
+linear according to the number of container elements as for linked lists). However, these
arrays are dynamic in the sense that they will automatically allocate more
memory if there is not enough of it for adding a new element. They also perform
range checking on the index values but in debug mode only, so please be sure to
The array classes were designed to be reasonably efficient, both in terms of
run-time speed and memory consumption and the executable size. The speed of
-array item access if, of course, constant (independent of number of elements)
+array item access is, of course, constant (independent of the number of elements)
making them much more efficient than linked lists (\helpref{wxList}{wxlist}).
Adding items to the arrays is also implemented in more or less constant time -
but the price is preallocating the memory in advance. In the \helpref{memory management}{wxarraymemorymanagement} section
wxArray is suitable for storing integer types and pointers which it does not
treat as objects in any way, i.e. the element pointed to by the pointer is not
-deleted when the element is removed from the array \&c. It should be noted that
-all of wxArray functions are inline, so it costs strictly nothing to define as
+deleted when the element is removed from the array. It should be noted that
+all of wxArray's functions are inline, so it costs strictly nothing to define as
many array types as you want (either in terms of the executable size or the
speed) as long as at least one of them is defined and this is always the case
-because wxArrays are used by wxWindows internally.
+because wxArrays are used by wxWindows internally. This class has one serious
+limitation: it can only be used for storing integral types (bool, char, short,
+int, long and their unsigned variants) or pointers (of any kind). An attempt
+to use with objects of sizeof() greater than sizeof(long) will provoke a
+runtime assertion failure, however declaring a wxArray of floats will not (on
+the machines where sizeof(float) <= sizeof(long)), yet it will {\bf not} work,
+please use wxObjArray for storing floats and doubles (NB: a more efficient
+wxArrayDouble class is scheduled for the next release of wxWindows).
wxSortedArray is a wxArray variant which should be used when searching in the
array is a frequently used operation. It requires you to define an additional
slower: it is $O(log(N))$ instead of constant time (neglecting time spent in
memory allocation routine). However, in a usual situation elements are added to
an array much less often than searched inside it, so wxSortedArray may lead to
-huge performance improvements compared to wxArray. As wxArray this array can not
-be used
+huge performance improvements compared to wxArray. Finally, it should be
+noticed that, as wxArray, wxSortedArray can be only used for storing integral
+types or pointers.
wxObjArray class treats its elements like "objects". It may delete them when
they are removed from the array (invoking the correct destructor) and copies
...
// now that we have MyDirectory declaration in scope we may finish the
-// definition of ArrayOfDirectories
+// definition of ArrayOfDirectories -- note that this expands into some C++
+// code and so should only be compiled once (i.e., don't put this in the
+// header, but into a source file or you will get linkin errors)
#include <wx/arrimpl.cpp> // this is a magic incantation which must be done!
WX_DEFINE_OBJARRAY(ArrayOfDirectories);
// that's all!
-
\end{verbatim}
It is not as elegant as writing
but is not that complicated and allows the code to be compiled with any, however
dumb, C++ compiler in the world.
-The things are much simpler for wxArray and wxSortedArray however: it is enough
+Things are much simpler for wxArray and wxSortedArray however: it is enough
just to write
\begin{verbatim}
\membersection{WX\_DEFINE\_ARRAY}\label{wxdefinearray}
-\func{}{WX\_DEFINE\_ARRAY}{\param{}{T}, \param{name}}
+\func{}{WX\_DEFINE\_ARRAY}{\param{}{T}, \param{}{name}}
This macro defines a new array class named {\it name} and containing the
elements of type {\it T}. Example:
the elements of type {\it T}. Example:
\begin{verbatim}
-WX_DEFINE_SORTED_ARRAY(int, wxArrayInt);
+WX_DEFINE_SORTED_ARRAY(int, wxSortedArrayInt);
class MyClass;
WX_DEFINE_SORTED_ARRAY(MyClass *, wxArrayOfMyClass);
return n1 - n2;
}
-wxArrayInt sorted(CompareInts);
+wxSortedArrayInt sorted(CompareInts);
int CompareMyClassObjects(MyClass *item1, MyClass *item2)
{
This macro declares a new object array class named {\it name} and containing
the elements of type {\it T}. Example:
+
\begin{verbatim}
class MyClass;
WX_DEFINE_OBJARRAY(MyClass, wxArrayOfMyClass); // note: not "MyClass *"!
\end{verbatim}
+
You must use \helpref{WX\_DEFINE\_OBJARRAY()}{wxdefineobjarray} macro to define
the array class - otherwise you would get link errors.
\membersection{WX\_CLEAR\_ARRAY}\label{wxcleararray}
-\func{\void}{WX\_CLEAR\_ARRAY}{\param{wxArray\& }{array}}
+\func{void}{WX\_CLEAR\_ARRAY}{\param{wxArray\& }{array}}
This macro may be used to delete all elements of the array before emptying it.
It can not be used with wxObjArrays - but they will delete their elements anyhow
when you call Empty().
-\membersection{Default constructors}\label{wxarrayctor}
+\membersection{Default constructors}\label{wxarrayctordef}
-\func{}{wxArray}{}
+\func{}{wxArray}{\void}
-\func{}{wxObjArray}{}
+\func{}{wxObjArray}{\void}
Default constructor initializes an empty array object.
should return a negative, zero or positive value according to whether the first
element passed to it is less than, equal to or greater than the second one.
-\membersection{wxArray copy constructor and assignemnt operator}\label{wxarrayctorcopy}
+\membersection{wxArray copy constructor and assignment operator}\label{wxarrayctorcopy}
\func{}{wxArray}{\param{const wxArray\& }{array}}
\func{}{wxObjArray}{\param{const wxObjArray\& }{array}}
-\func{wxArray\&}{operator=}{\param{const wxArray\& }{array}}
+\func{wxArray\&}{operator$=$}{\param{const wxArray\& }{array}}
-\func{wxSortedArray\&}{operator=}{\param{const wxSortedArray\& }{array}}
+\func{wxSortedArray\&}{operator$=$}{\param{const wxSortedArray\& }{array}}
-\func{wxObjArray\&}{operator=}{\param{const wxObjArray\& }{array}}
+\func{wxObjArray\&}{operator$=$}{\param{const wxObjArray\& }{array}}
The copy constructors and assignment operators perform a shallow array copy
(i.e. they don't copy the objects pointed to even if the source array contains
\membersection{wxArray::\destruct{wxArray}}\label{wxarraydtor}
-\func{}{\destruct{wxArray}}{}
+\func{}{\destruct{wxArray}}{\void}
-\func{}{\destruct{wxSortedArray}}{}
+\func{}{\destruct{wxSortedArray}}{\void}
-\func{}{\destruct{wxObjArray}}{}
+\func{}{\destruct{wxObjArray}}{\void}
The wxObjArray destructor deletes all the items owned by the array. This is not
done by wxArray and wxSortedArray versions - you may use
\membersection{wxArray::Add}\label{wxarrayadd}
-\func{\void}{Add}{\param{T }{item}}
+\func{void}{Add}{\param{T }{item}}
-\func{\void}{Add}{\param{T *}{item}}
+\func{void}{Add}{\param{T *}{item}}
-\func{\void}{Add}{\param{T \&}{item}}
+\func{void}{Add}{\param{T \&}{item}}
Appends a new element to the array (where {\it T} is the type of the array
elements.)
The first version is used with wxArray and wxSortedArray. The second and the
-third are used with wxObjArray. There is an {\bf important difference} between
+third are used with wxObjArray. There is an important difference between
them: if you give a pointer to the array, it will take ownership of it, i.e.
will delete it when the item is deleted from the array. If you give a reference
to the array, however, the array will make a copy of the item and will not take
\membersection{wxArray::Alloc}\label{wxarrayalloc}
-\func{\void}{Alloc}{\param{size\_t }{count}}
+\func{void}{Alloc}{\param{size\_t }{count}}
Preallocates memory for a given number of array elements. It is worth calling
when the number of items which are going to be added to the array is known in
\membersection{wxArray::Clear}\label{wxarrayclear}
-\func{\void}{Clear}{\void}
+\func{void}{Clear}{\void}
This function does the same as \helpref{Empty()}{wxarrayempty} and additionally
frees the memory allocated to the array.
\membersection{wxArray::Empty}\label{wxarrayempty}
-\func{\void}{Empty}{\void}
+\func{void}{Empty}{\void}
Empties the array. For wxObjArray classes, this destroys all of the array
elements. For wxArray and wxSortedArray this does nothing except marking the
-array of being empty - this function does not free the allocated memory, use
+array of being empty - this function does not free the allocated memory, use
\helpref{Clear()}{wxarrayclear} for this.
\membersection{wxArray::GetCount}\label{wxarraygetcount}
\membersection{wxArray::Insert}\label{wxarrayinsert}
-\func{\void}{Insert}{\param{T }{item}, \param{size\_t }{n}}
+\func{void}{Insert}{\param{T }{item}, \param{size\_t }{n}}
-\func{\void}{Insert}{\param{T *}{item}, \param{size\_t }{n}}
+\func{void}{Insert}{\param{T *}{item}, \param{size\_t }{n}}
-\func{\void}{Insert}{\param{T \&}{item}, \param{size\_t }{n}}
+\func{void}{Insert}{\param{T \&}{item}, \param{size\_t }{n}}
-Insert a new item into the array before the item {\it n} - thus, {\it
-Insert(something, 0u}} will insert an item in such way that it will become the
+Insert a new item into the array before the item {\it n} - thus, {\it Insert(something, 0u)} will
+insert an item in such way that it will become the
first array element.
Please see \helpref{Add()}{wxarrayadd} for explanation of the differences
\membersection{wxArray::IsEmpty}\label{wxarrayisempty}
-\constfunc{bool}{IsEmpty}{}
+\constfunc{bool}{IsEmpty}{\void}
Returns TRUE if the array is empty, FALSE otherwise.
\membersection{wxArray::Shrink}\label{wxarrayshrink}
-\func{\void}{Shrink}{\void}
+\func{void}{Shrink}{\void}
Frees all memory unused by the array. If the program knows that no new items
will be added to the array it may call Shrink() to reduce its memory usage.
\membersection{wxArray::Sort}\label{wxarraysort}
-\func{\void}{Sort}{\param{CMPFUNC<T> }{compareFunction}}
+\func{void}{Sort}{\param{CMPFUNC<T> }{compareFunction}}
The notation CMPFUNC<T> should be read as if we had the following declaration:
template int CMPFUNC(T *first, T *second);
\end{verbatim}
-where {\it T} is the type of the array elements. I.e. it is a function returning
+where {\it T} is the type of the array elements. I.e. it is a function returning
{\it int} which is passed two arguments of type {\it T *}.
Sorts the array using the specified compare function: this function should
projects / wxWidgets.git / blobdiff